INDEX

Explanations

personal pronouns followed by verbs of communication like "ask", "tell", or "teach"

oai_token-act-pair · gpt-3.5-turbo

references to people being asked questions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ourning

-0.77

Ĥ¬

-0.76

acca

-0.76

audio

-0.66

BuyableInstoreAndOnline

-0.66

aten

-0.66

pite

-0.65

kered

-0.65

ushima

-0.63

edition

-0.62

POSITIVE LOGITS

 questions

1.12

 rhet

1.00

 probing

0.95

 question

0.92

 forgiveness

0.90

 QUEST

0.87

 Questions

0.83

 politely

0.82

 plaint

0.82

question

0.82

Activations Density 0.120%

personal pronouns followed by verbs of communication like "ask", "tell", or "teach"

references to people being asked questions

No Comments

No Known Activations