INDEX

Explanations

names of political figures or organizations

oai_token-act-pair · gpt-3.5-turbo

references to former officials and their positions

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

itivity

-0.69

ickle

-0.69

shapeshifter

-0.68

ptoms

-0.66

ignant

-0.66

ebook

-0.65

minecraft

-0.64

ankind

-0.64

oeuv

-0.62

itars

-0.61

POSITIVE LOGITS

£

0.78

ģĸ

0.77

 operative

0.75

¢

0.74

 stal

0.74

į

0.72

oslov

0.68

ãĤ¶

0.68

§

0.68

 Olympia

0.68

Activations Density 0.167%

names of political figures or organizations

references to former officials and their positions

No Comments

No Known Activations