INDEX

Explanations

phrases indicating strong positive or negative opinions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a positive assessment or recognition of quality

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

pez

-0.66

hyde

-0.66

rely

-0.66

yahoo

-0.65

oleon

-0.63

ggle

-0.63

omore

-0.63

inity

-0.60

 Samp

-0.60

iferation

-0.60

POSITIVE LOGITS

 suited

1.10

spring

1.08

 vers

1.04

 behaved

1.03

 enough

0.99

enough

0.97

 acquainted

0.96

 positioned

0.95

 documented

0.94

 regarded

0.94

Activations Density 0.035%

phrases indicating strong positive or negative opinions

phrases indicating a positive assessment or recognition of quality

No Comments

No Known Activations