INDEX

Explanations

positive adjectives describing something pleasant or agreeable

oai_token-act-pair · gpt-3.5-turbo

the word "nice" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

aer

-0.90

arian

-0.84

rained

-0.80

mental

-0.78

arians

-0.77

avis

-0.72

ipt

-0.67

inant

-0.67

produced

-0.66

Hop

-0.66

POSITIVE LOGITS

 nice

1.03

nice

0.84

bye

0.83

 nifty

0.82

 enough

0.81

 reception

0.77

 additions

0.76

 bonus

0.75

 decent

0.75

 neat

0.73

Activations Density 0.012%

positive adjectives describing something pleasant or agreeable

the word "nice" in various contexts

No Comments

No Known Activations