INDEX

Explanations

mentions of legal or regulatory restrictions

oai_token-act-pair · gpt-3.5-turbo

terms related to limitations or controls imposed on actions or freedoms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

story

-0.75

tein

-0.75

Sea

-0.72

past

-0.72

 Lakes

-0.71

rious

-0.70

nie

-0.70

sis

-0.70

 Chemistry

-0.68

Hy

-0.67

POSITIVE LOGITS

 restrictions

1.40

 restricting

1.22

 imposed

1.16

 restriction

1.10

 prohibited

1.09

 restricts

1.05

 prohibitions

1.02

 exemptions

0.99

 restrictive

0.97

 restricted

0.96

Activations Density 0.008%

mentions of legal or regulatory restrictions

terms related to limitations or controls imposed on actions or freedoms

No Comments

No Known Activations