INDEX

Explanations

mentions of names or terms related to politicians or political figures, specifically focusing on the term "iden"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals' names

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

 exha

-0.82

thora

-0.81

 Canadians

-0.70

 Rita

-0.69

 uncontrolled

-0.69

 weeds

-0.67

 challeng

-0.65

ãħĭ

-0.63

pmwiki

-0.63

Canadian

-0.62

POSITIVE LOGITS

iden

1.47

ovo

0.90

unci

0.89

ners

0.87

ning

0.86

ned

0.85

ception

0.85

roid

0.77

nen

0.76

vier

0.76

Activations Density 0.006%

mentions of names or terms related to politicians or political figures, specifically focusing on the term "iden"

references to specific individuals' names

No Comments

No Known Activations