INDEX

Explanations

mentions of historical figures and events

oai_token-act-pair · gpt-3.5-turbo

references to historical texts and accounts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CEO

-0.81

 ObamaCare

-0.73

 defund

-0.72

 STOP

-0.72

NYSE

-0.71

afety

-0.70

 treadmill

-0.70

 Bravo

-0.69

 Obamacare

-0.69

oppable

-0.69

POSITIVE LOGITS

 extant

1.20

 historians

1.20

 Herod

1.19

 historian

1.18

 manuscripts

1.17

 histor

1.12

 interpol

1.09

 mentions

1.06

 Hist

1.05

 inscription

1.04

Activations Density 0.448%

mentions of historical figures and events

references to historical texts and accounts

No Comments

No Known Activations