INDEX

Explanations

names of specific locations or institutions

oai_token-act-pair · gpt-3.5-turbo

words associated with extreme events or conditions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Vaugh

-0.58

 prest

-0.58

 referen

-0.56

 corrid

-0.54

 thous

-0.53

sic

-0.52

_.

-0.52

 destro

-0.52

 disadvant

-0.51

 challeng

-0.51

POSITIVE LOGITS

 Kingdoms

0.59

 Rewards

0.53

 GOODMAN

0.51

 Profile

0.51

 Hedge

0.51

Streamer

0.51

 Seasons

0.50

 âĢº

0.50

 Scene

0.49

 Shin

0.49

Activations Density 1.753%

names of specific locations or institutions

words associated with extreme events or conditions

No Comments

No Known Activations