INDEX

Explanations

statements indicating justification or lack of justification for actions or decisions made

oai_token-act-pair · gpt-3.5-turbo

instances of the word "justified" and related terms expressing legitimacy or reasonableness

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

pmwiki

-0.83

ammy

-0.80

thin

-0.73

hm

-0.73

berry

-0.69

ept

-0.69

apple

-0.67

 Huck

-0.67

stone

-0.66

ummer

-0.65

POSITIVE LOGITS

 justified

1.07

 justifies

1.05

 justification

0.81

 indignation

0.78

 guiActiveUn

0.76

ifiable

0.76

 justify

0.76

ensibly

0.72

 compe

0.70

 justifying

0.70

Activations Density 0.009%

statements indicating justification or lack of justification for actions or decisions made

instances of the word "justified" and related terms expressing legitimacy or reasonableness

No Comments

No Known Activations