INDEX

Explanations

phrases related to actions or scenarios that can be done or experienced without certain conditions or constraints

oai_token-act-pair · gpt-3.5-turbo

phrases indicating absence or lack of something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

oola

-0.91

mun

-0.83

raq

-0.76

omsky

-0.75

ahime

-0.74

ocations

-0.68

ricks

-0.67

ais

-0.67

lish

-0.66

des

-0.66

POSITIVE LOGITS

lessly

0.86

 knowing

0.81

 blinking

0.75

 sacrificing

0.71

 regard

0.70

 compromising

0.69

etheless

0.68

 touching

0.67

 forgetting

0.66

 substit

0.66

Activations Density 0.027%

phrases related to actions or scenarios that can be done or experienced without certain conditions or constraints

phrases indicating absence or lack of something

No Comments

No Known Activations