INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper names and characters in the text

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

committee

-0.75

 industrial

-0.71

 antim

-0.71

rador

-0.70

£ı

-0.67

flat

-0.66

 disinfect

-0.65

Engineers

-0.64

inline

-0.64

ffic

-0.62

POSITIVE LOGITS

 Jenner

0.96

Doe

0.93

 herself

0.93

's

0.89

Kir

0.84

 Weasley

0.82

Mae

0.82

Lup

0.81

 Rivera

0.79

 Gund

0.79

Activations Density 0.169%

names of individuals

proper names and characters in the text

No Comments

No Known Activations