INDEX

Explanations

phrases related to recognition or praise

oai_token-act-pair · gpt-3.5-turbo

expressions of recognition or commendation for individuals or groups

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

cum

-0.85

opic

-0.80

hire

-0.79

credit

-0.77

eenth

-0.75

bang

-0.74

lust

-0.71

chie

-0.69

wise

-0.68

gas

-0.66

POSITIVE LOGITS

Reviewer

0.76

 honored

0.76

 DeVos

0.75

 honoured

0.68

nesday

0.68

 guests

0.68

 Parish

0.64

ĸļ

0.63

âĿ

0.62

 toast

0.62

Activations Density 0.016%

phrases related to recognition or praise

expressions of recognition or commendation for individuals or groups

No Comments

No Known Activations