INDEX

Explanations

mentions of actions performed or decisions made by someone other than the subject

oai_token-act-pair · gpt-3.5-turbo

references to accountability or actions taken by others

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

olid

-0.73

undown

-0.72

tnc

-0.72

rar

-0.71

Accessory

-0.67

Increases

-0.66

veyard

-0.60

osterone

-0.60

taining

-0.60

irth

-0.60

POSITIVE LOGITS

 whom

0.84

who

0.83

worldly

0.81

's

0.75

 chooses

0.73

ÃŃs

0.73

 chose

0.72

zzi

0.68

 fared

0.68

 thinks

0.67

Activations Density 0.152%

mentions of actions performed or decisions made by someone other than the subject

references to accountability or actions taken by others

No Comments

No Known Activations