INDEX

Explanations

phrases indicating uncertainty or incomplete information

oai_token-act-pair · gpt-3.5-turbo

the word "yet" indicating uncertainty or incompleteness of a situation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ufact

-0.75

rities

-0.71

gang

-0.69

-0.66

 grievance

-0.66

mut

-0.65

case

-0.65

gencies

-0.64

ospel

-0.62

cro

-0.61

POSITIVE LOGITS

hin

0.86

abouts

0.75

terday

0.74

uden

0.69

 anyway

0.68

adays

0.68

agos

0.67

èª

0.65

 officially

0.64

aquin

0.64

Activations Density 0.029%

phrases indicating uncertainty or incomplete information

the word "yet" indicating uncertainty or incompleteness of a situation

No Comments

No Known Activations