INDEX

Explanations

questions or statements with the word "how"

oai_token-act-pair · gpt-3.5-turbo

queries or discussions about uncertainty and the manner of various situations or events

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

ibl

-0.67

ubs

-0.66

gin

-0.66

usa

-0.64

agin

-0.64

sic

-0.63

yon

-0.63

athering

-0.62

article

-0.62

inas

-0.61

POSITIVE LOGITS

 much

1.33

far

1.16

 many

1.13

 long

1.13

 badly

1.06

much

1.04

 often

1.04

 closely

1.03

 quickly

1.03

 prevalent

1.01

Activations Density 0.071%

questions or statements with the word "how"

queries or discussions about uncertainty and the manner of various situations or events

No Comments

No Known Activations