INDEX

Explanations

pronouns followed by verbs

oai_token-act-pair · gpt-3.5-turbo

references to pronouns that indicate subjects involved in actions or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

yss

-0.76

aughtered

-0.73

hatt

-0.72

 Barron

-0.69

 Berkshire

-0.69

erala

-0.66

rets

-0.64

abel

-0.64

retty

-0.63

 Dres

-0.63

POSITIVE LOGITS

'd

0.87

 proport

0.76

should

0.74

 deems

0.74

'll

0.73

 might

0.73

might

0.73

ģĸ

0.72

 could

0.70

self

0.69

Activations Density 0.217%

pronouns followed by verbs

references to pronouns that indicate subjects involved in actions or situations

No Comments

No Known Activations