INDEX

Explanations

negative or challenging situations

oai_token-act-pair · gpt-3.5-turbo

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

ased

-0.67

tein

-0.67

hai

-0.64

yan

-0.61

acht

-0.60

lass

-0.60

ase

-0.60

TW

-0.59

ocard

-0.58

 Mant

-0.57

POSITIVE LOGITS

any

1.06

 slightest

1.04

 anything

1.03

 ever

1.02

 anyone

0.97

 anybody

0.96

 whatsoever

0.88

 anywhere

0.86

 mention

0.86

 remotely

0.84

Activations Density 0.673%

negative or challenging situations

No Comments

No Known Activations