INDEX

Explanations

phrases indicating the act of revealing information or oneself

oai_token-act-pair · gpt-3.5-turbo

instances of revealing information or disclosures related to personal or confidential matters

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

Reviewed

-0.74

oslav

-0.74

 recognizes

-0.64

 emulate

-0.64

hesda

-0.62

chairs

-0.62

 realize

-0.60

 upkeep

-0.60

rollers

-0.59

eson

-0.59

POSITIVE LOGITS

 secrets

0.97

 clues

0.85

 mysteries

0.80

 trove

0.78

 WikiLeaks

0.75

 vulnerabilities

0.73

 whereabouts

0.72

İĭ

0.72

 incrim

0.72

 truths

0.70

Activations Density 0.252%

phrases indicating the act of revealing information or oneself

instances of revealing information or disclosures related to personal or confidential matters

No Comments

No Known Activations