INDEX

Explanations

descriptions related to spoken or written expressions

oai_token-act-pair · gpt-3.5-turbo

references to speeches

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

olen

-0.73

ETA

-0.70

VAL

-0.65

 Cron

-0.64

Ã±o

-0.64

astern

-0.64

nerg

-0.61

involved

-0.60

 Patri

-0.60

nergy

-0.60

POSITIVE LOGITS

 speech

1.31

speech

1.21

 Speech

1.13

 speeches

0.93

anooga

0.93

heim

0.89

spe

0.80

 speaker

0.79

voice

0.79

 onstage

0.78

Activations Density 0.014%

descriptions related to spoken or written expressions

references to speeches

No Comments

No Known Activations