INDEX

Explanations

mentions of innocence and innocence-related terms

oai_token-act-pair · gpt-3.5-turbo

references to innocence or innocent individuals

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orders

-0.79

anwhile

-0.77

Recomm

-0.76

division

-0.74

atche

-0.72

KEY

-0.69

ridor

-0.68

åĭ

-0.67

ingo

-0.67

artney

-0.67

POSITIVE LOGITS

 innocent

1.43

 innocence

1.22

 bystand

1.09

 bystanders

0.98

ocent

0.93

 innoc

0.84

 Innocent

0.79

 civilians

0.73

 harmless

0.70

 mole

0.69

Activations Density 0.010%

mentions of innocence and innocence-related terms

references to innocence or innocent individuals

No Comments

No Known Activations