INDEX

Explanations

instances of the word "admit"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "admit" 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 @ 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

rior

-0.85

chn

-0.76

ccording

-0.71

nton

-0.69

locked

-0.68

ammy

-0.68

tsky

-0.68

miah

-0.66

nc

-0.66

ighth

-0.64

POSITIVE LOGITS

ibility

0.92

 admit

0.86

 defeat

0.82

 wrongdoing

0.80

iary

0.79

IBLE

0.78

 admitted

0.76

orial

0.74

 admitting

0.74

essors

0.73

Activations Density 0.018%

instances of the word "admit"

instances of the word "admit" in various contexts

No Comments

No Known Activations