INDEX

Explanations

active verbs related to creating, adjusting, or controlling something

oai_token-act-pair · gpt-3.5-turbo

commands or intentions related to creating or modifying something

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wikipedia

-0.70

wordpress

-0.62

amm

-0.62

 Subtle

-0.58

 Hurt

-0.57

oggles

-0.57

ãģ®

-0.57

 Dresden

-0.56

 modeled

-0.56

 Heck

-0.55

POSITIVE LOGITS

().

0.75

 them

0.71

uate

0.69

 Ratio

0.68

hy

0.66

itial

0.64

 compliance

0.64

 purposes

0.63

agate

0.62

peak

0.62

Activations Density 0.341%

active verbs related to creating, adjusting, or controlling something

commands or intentions related to creating or modifying something

No Comments

No Known Activations