INDEX

Explanations

instances where something is being allowed or permitted

oai_token-act-pair · gpt-3.5-turbo

the word "let" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 holiest

-0.87

 Soldier

-0.70

ONSORED

-0.69

nown

-0.65

 Languages

-0.64

PLIED

-0.62

 Mouth

-0.62

atana

-0.59

 favour

-0.59

Zen

-0.59

POSITIVE LOGITS

icia

0.85

staking

0.84

tered

0.82

us

0.77

 lett

0.74

tering

0.73

 slip

0.73

let

0.71

¶ħ

0.68

winter

0.68

Activations Density 0.028%

instances where something is being allowed or permitted

the word "let" in various contexts

No Comments

No Known Activations