INDEX

Explanations

phrases related to consecutive occurrences or repeated actions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating repetition over multiple years

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

merce

-0.94

 mathemat

-0.77

ascript

-0.75

 metic

-0.73

omorphic

-0.72

ufact

-0.70

 streng

-0.68

 competence

-0.67

urate

-0.66

 mosqu

-0.66

POSITIVE LOGITS

dy

1.19

dies

1.05

row

1.00

Row

0.87

er

0.85

ser

0.85

feed

0.81

 rows

0.78

boat

0.77

dale

0.75

Activations Density 0.006%

phrases related to consecutive occurrences or repeated actions

phrases indicating repetition over multiple years

No Comments

No Known Activations