INDEX

Explanations

mentions of innocence and associated concepts

oai_token-act-pair · gpt-3.5-turbo

mentions of innocent individuals or groups

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

division

-0.74

TOP

-0.74

sych

-0.73

ANN

-0.72

artney

-0.71

jri

-0.71

riber

-0.69

ingo

-0.69

pain

-0.69

è¦ļéĨĴ

-0.68

POSITIVE LOGITS

 bystand

1.37

 bystanders

1.29

 innocent

1.10

 innocence

0.91

 civilians

0.84

 victims

0.84

 prey

0.81

ocent

0.81

 innoc

0.77

ously

0.73

Activations Density 0.017%

mentions of innocence and associated concepts

mentions of innocent individuals or groups

No Comments

No Known Activations