INDEX

Explanations

phrases related to general topics or concepts

oai_token-act-pair · gpt-3.5-turbo

references to generalized concepts or items described as "things."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

inav

-0.61

avorite

-0.60

KM

-0.57

cul

-0.56

 crest

-0.55

ynski

-0.54

ÂłÂłÂłÂłÂłÂłÂłÂł

-0.54

 commentary

-0.53

ped

-0.53

onz

-0.52

POSITIVE LOGITS

iverse

1.36

 happened

1.03

 happening

1.00

 happen

0.88

ional

0.88

 happens

0.85

Else

0.83

hots

0.78

ies

0.77

 happ

0.75

Activations Density 0.063%

phrases related to general topics or concepts

references to generalized concepts or items described as "things."

No Comments

No Known Activations