INDEX

Explanations

words related to a strong sense of wonder, admiration, or respect

oai_token-act-pair · gpt-3.5-turbo

expressions of admiration and emotional responses

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

 Thieves

-0.69

sol

-0.69

å¥³

-0.64

Interstitial

-0.64

Short

-0.61

Opt

-0.61

dule

-0.60

 Noir

-0.60

 Agents

-0.60

 Obesity

-0.59

POSITIVE LOGITS

awe

0.99

urous

0.98

ingly

0.95

eful

0.90

efully

0.87

htaking

0.83

 aback

0.75

 amaz

0.75

fulness

0.74

worthy

0.72

Activations Density 0.015%

words related to a strong sense of wonder, admiration, or respect

expressions of admiration and emotional responses

No Comments

No Known Activations