INDEX

Explanations

phrases related to step-by-step processes

oai_token-act-pair · gpt-3.5-turbo

references to sequential processes or instructions

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

ãĥĩãĤ£

-0.82

 Predator

-0.76

luaj

-0.66

ucc

-0.63

 scraps

-0.61

Owl

-0.60

United

-0.60

 occupant

-0.60

MEN

-0.59

cov

-0.59

POSITIVE LOGITS

udic

0.87

steps

0.85

 steps

0.79

 step

0.69

 Steps

0.69

antry

0.68

 Step

0.65

 directions

0.64

oward

0.64

frog

0.62

Activations Density 0.073%

phrases related to step-by-step processes

references to sequential processes or instructions

No Comments

No Known Activations