INDEX

Explanations

phrases related to changes or transformations

oai_token-act-pair · gpt-3.5-turbo

phrases and terms related to escalating situations or conflicts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MQ

-0.62

 Supports

-0.62

 Instruction

-0.58

thia

-0.57

 cabinets

-0.57

stall

-0.56

 bowel

-0.56

 Releases

-0.56

ools

-0.55

 dumps

-0.54

POSITIVE LOGITS

ĸļ

0.66

iHUD

0.63

\\\\\\\\

0.61

 unden

0.59

 borne

0.59

hidden

0.57

VPN

0.57

 Hond

0.57

------------------------------------------------

0.57

FontSize

0.57

Activations Density 1.048%

phrases related to changes or transformations

phrases and terms related to escalating situations or conflicts

No Comments

No Known Activations