INDEX

Explanations

references to policy-related words and phrases

oai_token-act-pair · gpt-3.5-turbo

references to policy-related organizations and studies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 captcha

-0.82

ysis

-0.78

yang

-0.72

ursion

-0.70

enance

-0.68

ãĤ¨ãĥ«

-0.68

rises

-0.67

 dimensional

-0.66

bley

-0.66

 recall

-0.65

POSITIVE LOGITS

 Agency

0.96

 Institute

0.89

 Accountability

0.88

 Forum

0.86

 Analysis

0.86

edia

0.85

 Fund

0.83

 Analyst

0.83

 Strategies

0.82

 Blog

0.82

Activations Density 0.034%

references to policy-related words and phrases

references to policy-related organizations and studies

No Comments

No Known Activations