INDEX

Explanations

numbers followed by a colon indicating a list item

oai_token-act-pair · gpt-3.5-turbo

numerical scores and rankings associated with specific entities or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

mble

-0.77

 arrang

-0.71

chwitz

-0.70

 citiz

-0.69

 agre

-0.69

objects

-0.68

yond

-0.64

soType

-0.64

uve

-0.63

intent

-0.62

POSITIVE LOGITS

 Prem

0.62

 Lack

0.62

 Hilbert

0.62

 Myth

0.62

odcast

0.61

 Introdu

0.60

PRO

0.60

TBD

0.59

LEG

0.58

æ©Ł

0.58

Activations Density 0.195%

numbers followed by a colon indicating a list item

numerical scores and rankings associated with specific entities or events

No Comments

No Known Activations