INDEX

Explanations

phrases related to conspiracy or collusion

oai_token-act-pair · gpt-3.5-turbo

mentions of "cons" or related terms suggesting some form of conspiracy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

baugh

-0.69

Lup

-0.61

 Flint

-0.60

 Forrest

-0.60

 Highlander

-0.60

 Hills

-0.59

 streams

-0.59

WOOD

-0.59

 oats

-0.58

 withdrawals

-0.58

POSITIVE LOGITS

cientious

0.97

 cons

0.93

cription

0.92

ignment

0.89

oler

0.88

afety

0.88

ensual

0.88

oled

0.87

erver

0.86

igned

0.85

Activations Density 0.005%

phrases related to conspiracy or collusion

mentions of "cons" or related terms suggesting some form of conspiracy

No Comments

No Known Activations