INDEX

Explanations

personal pronouns followed by a verb indicating action being done to someone or something

oai_token-act-pair · gpt-3.5-turbo

references to individuals or groups being left in vulnerable situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

GW

-0.71

wcsstore

-0.69

 Cosponsors

-0.68

idth

-0.65

anta

-0.64

fixes

-0.62

haps

-0.59

usual

-0.58

 Drill

-0.58

afort

-0.58

POSITIVE LOGITS

 untouched

1.32

 intact

1.28

 behind

1.12

 unchanged

1.12

 unatt

1.10

 unprotected

1.07

 unaffected

1.07

 stranded

1.06

 dangling

1.05

 wondering

1.04

Activations Density 0.114%

personal pronouns followed by a verb indicating action being done to someone or something

references to individuals or groups being left in vulnerable situations

No Comments

No Known Activations