INDEX

Explanations

terms related to interactions or processes that are engaging, interesting, or draw attention

oai_token-act-pair · gpt-3.5-turbo

variations of the word "extract."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

racuse

-0.69

 Sons

-0.67

 laundering

-0.62

swick

-0.60

 slideshow

-0.59

viron

-0.59

heit

-0.59

lap

-0.59

kees

-0.58

steen

-0.58

POSITIVE LOGITS

ract

1.53

ractor

1.05

ors

0.96

ress

0.94

racted

0.92

ic

0.90

ict

0.90

icular

0.90

inct

0.84

ory

0.84

Activations Density 0.010%

terms related to interactions or processes that are engaging, interesting, or draw attention

variations of the word "extract."

No Comments

No Known Activations