INDEX

Explanations

words related to sets, collections, or groups

oai_token-act-pair · gpt-3.5-turbo

references to technical standards or guidelines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

proof

-0.63

 reluctance

-0.60

 parent

-0.60

 jewel

-0.57

 subdu

-0.57

 ruling

-0.56

anyl

-0.56

 horror

-0.56

 injustice

-0.55

mum

-0.55

POSITIVE LOGITS

ets

4.73

ETS

2.93

eting

2.30

et

2.26

ET

1.93

etting

1.83

etry

1.81

etts

1.71

etary

1.69

eters

1.66

Activations Density 0.007%

words related to sets, collections, or groups

references to technical standards or guidelines

No Comments

No Known Activations