INDEX

Explanations

references to the location "Topeka"

oai_token-act-pair · gpt-3.5-turbo

references to the location "Topeka" and mentions of "Kobe" related to sneakers or taxes

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Hallow

-0.80

 glim

-0.77

ingham

-0.76

yss

-0.75

ngth

-0.75

 challeng

-0.71

orget

-0.71

aire

-0.66

 Lauder

-0.66

ettings

-0.66

POSITIVE LOGITS

ishi

0.83

RECT

0.74

rica

0.73

raham

0.71

els

0.70

jured

0.70

amel

0.70

sole

0.70

ceivable

0.69

lesi

0.69

Activations Density 0.027%

references to the location "Topeka"

references to the location "Topeka" and mentions of "Kobe" related to sneakers or taxes

No Comments

No Known Activations