INDEX

Explanations

references to witnesses

oai_token-act-pair · gpt-3.5-turbo

references to witnesses 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

arter

-0.79

elsius

-0.79

ogo

-0.73

 nonviolent

-0.71

ofi

-0.70

isine

-0.70

 flush

-0.70

 evenly

-0.69

 chill

-0.66

ono

-0.65

POSITIVE LOGITS

 witness

1.26

 witnesses

1.19

Witness

1.18

 testimonies

1.10

 Witness

1.04

 testimony

0.98

ewitness

0.91

 thereto

0.84

 Witnesses

0.81

wcsstore

0.77

Activations Density 0.008%

references to witnesses

references to witnesses in various contexts

No Comments

No Known Activations