INDEX

Explanations

phrases related to negative criticism or attacks

oai_token-act-pair · gpt-3.5-turbo

instances of the word "belittle" and its variations

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

inals

-0.79

office

-0.79

 Fields

-0.72

anwhile

-0.71

INAL

-0.70

hibition

-0.69

 Enhancement

-0.68

 Regulation

-0.68

exempt

-0.67

azines

-0.65

POSITIVE LOGITS

ittle

0.90

gian

0.87

bel

0.84

aying

0.83

ayed

0.82

ieving

0.79

ayer

0.78

phe

0.76

ieved

0.74

ousing

0.73

Activations Density 0.010%

phrases related to negative criticism or attacks

instances of the word "belittle" and its variations

No Comments

No Known Activations