INDEX

Explanations

words related to political entities like parties and associations

oai_token-act-pair · gpt-3.5-turbo

sentences that have a significant punctuation or structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 guiActiveUn

-0.81

 advis

-0.78

wagen

-0.76

 stewards

-0.75

 withd

-0.71

ihar

-0.70

 portrait

-0.69

 dictate

-0.68

 volunte

-0.67

 charact

-0.67

POSITIVE LOGITS

org

0.78

0.76

Org

0.76

jpg

0.75

com

0.74

Va

0.73

wikipedia

0.73

 coli

0.73

asc

0.72

0.71

Activations Density 0.129%

words related to political entities like parties and associations

sentences that have a significant punctuation or structure

No Comments

No Known Activations