INDEX

Explanations

references to physical locations or landmarks

oai_token-act-pair · gpt-3.5-turbo

mentions of "crossroads" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Ĥ

-0.72

¶ħ

-0.71

 quir

-0.70

Ĥİ

-0.69

ulous

-0.64

ever

-0.63

Ģ

-0.63

 basic

-0.62

 judges

-0.62

ishable

-0.61

POSITIVE LOGITS

roads

1.19

NetMessage

0.97

 Roads

0.90

ãĤ¤ãĥĪ

0.80

PsyNetMessage

0.74

road

0.74

 Showdown

0.70

FTWARE

0.70

lisher

0.70

bourg

0.68

Activations Density 0.020%

references to physical locations or landmarks

mentions of "crossroads" in various contexts

No Comments

No Known Activations