INDEX

Explanations

phrases indicating specific actions or events related to information dissemination or updates

oai_token-act-pair · gpt-3.5-turbo

actions that imply ongoing changes or updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

arist

-0.65

rouse

-0.64

ears

-0.64

 Latter

-0.61

article

-0.61

rients

-0.61

chal

-0.57

dos

-0.55

cp

-0.54

 greets

-0.54

POSITIVE LOGITS

 lately

0.79

 since

0.74

IFIED

0.73

 psychiat

0.70

Ú

0.70

opted

0.69

iasco

0.69

kef

0.69

 unanimously

0.68

Ļ½

0.68

Activations Density 0.154%

phrases indicating specific actions or events related to information dissemination or updates

actions that imply ongoing changes or updates

No Comments

No Known Activations