INDEX

Explanations

references to specific events or actions related to individuals

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

phrases that indicate popularity or reception, particularly in the context of media or performance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

arta

-0.82

-|

-0.77

ACH

-0.75

querque

-0.75

rue

-0.74

asking

-0.74

aby

-0.72

ondo

-0.72

 Alone

-0.71

Gas

-0.71

POSITIVE LOGITS

 rave

1.34

 widespread

1.27

 praise

1.15

 considerable

1.14

 notoriety

1.12

 acclaim

1.08

 glowing

1.04

 accol

1.04

 scant

1.02

 controversy

1.02

Activations Density 0.202%

references to specific events or actions related to individuals

phrases that indicate popularity or reception, particularly in the context of media or performance

No Comments

No Known Activations