INDEX

Explanations

references to the app store, specifically Google Play

oai_token-act-pair · gpt-3.5-turbo

references to the Google Play Store

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Beir

-0.88

pora

-0.69

 pole

-0.66

 leap

-0.65

zzle

-0.65

 deviation

-0.63

 detachment

-0.63

 fumes

-0.63

 dilig

-0.62

 hemorrh

-0.62

POSITIVE LOGITS

wright

1.20

ground

1.15

GROUND

1.13

ername

1.09

grounds

1.07

lists

1.05

dates

1.04

station

0.96

mates

0.94

testers

0.92

Activations Density 0.034%

references to the app store, specifically Google Play

references to the Google Play Store

No Comments

No Known Activations