INDEX

Explanations

phrases or sentences that transition to a new topic or idea

oai_token-act-pair · gpt-3.5-turbo

phrases that suggest transitions or conclusions in discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

iership

-0.65

agos

-0.64

ricular

-0.64

 destro

-0.63

 withdrew

-0.60

 financed

-0.59

roman

-0.59

 tolerated

-0.58

 cour

-0.58

 vanquished

-0.57

POSITIVE LOGITS

me

0.89

 curious

0.80

undrum

0.78

 unsur

0.77

 intriguing

0.77

 doub

0.76

 fascinating

0.76

 begs

0.76

question

0.74

 obvious

0.74

Activations Density 0.138%

phrases or sentences that transition to a new topic or idea

phrases that suggest transitions or conclusions in discussions

No Comments

No Known Activations