INDEX

Explanations

actions of preventing something or someone from a specific activity or location

oai_token-act-pair · gpt-3.5-turbo

phrases related to actions or processes that involve preventing someone or something from taking a specific action

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

clair

-0.70

çīĪ

-0.70

committee

-0.67

 headed

-0.64

dimension

-0.64

mac

-0.64

 summar

-0.63

hematic

-0.63

 doubtless

-0.61

eria

-0.60

POSITIVE LOGITS

 anymore

0.88

 anything

0.87

 altogether

0.82

BuyableInstoreAndOnline

0.76

 harmful

0.74

any

0.74

nor

0.70

 unsafe

0.67

ratch

0.66

 whatsoever

0.65

Activations Density 0.156%

actions of preventing something or someone from a specific activity or location

phrases related to actions or processes that involve preventing someone or something from taking a specific action

No Comments

No Known Activations