INDEX

Explanations

phrases expressing strong opinions or evaluations

oai_token-act-pair · gpt-3.5-turbo

statements that assess or critique the validity or morality of opinions or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 specialize

-0.81

atile

-0.75

 reorgan

-0.71

 slic

-0.69

 retrie

-0.66

 migrate

-0.66

 tables

-0.65

 shuttle

-0.65

 roam

-0.65

athered

-0.65

POSITIVE LOGITS

udicrous

0.86

quickShipAvailable

0.85

 ludicrous

0.84

Absolutely

0.82

 understandable

0.82

 laughable

0.81

SourceFile

0.80

Probably

0.79

 understatement

0.79

 disingen

0.79

Activations Density 0.289%

phrases expressing strong opinions or evaluations

statements that assess or critique the validity or morality of opinions or situations

No Comments

No Known Activations