INDEX

Explanations

words related to labeling or categorizing

oai_token-act-pair · gpt-3.5-turbo

terms related to categorization and labeling

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

perty

-0.85

rentice

-0.69

 balance

-0.64

=-=-=-=-

-0.62

elin

-0.62

 ende

-0.61

vous

-0.61

 attendant

-0.60

 purse

-0.60

unia

-0.60

POSITIVE LOGITS

phas

0.86

ifferent

0.77

ostic

0.77

icut

0.76

ifiable

0.73

ifications

0.72

enance

0.72

ifier

0.72

oma

0.71

naire

0.70

Activations Density 0.105%

words related to labeling or categorizing

terms related to categorization and labeling

No Comments

No Known Activations