INDEX

Explanations

patterns related to events happening immediately one after another

oai_token-act-pair · gpt-3.5-turbo

phrases indicating sequential events or occurrences

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

kept

-0.65

 unfocusedRange

-0.65

enberg

-0.64

oret

-0.62

oS

-0.62

utenberg

-0.61

oire

-0.61

acles

-0.60

Ble

-0.60

database

-0.60

POSITIVE LOGITS

noon

0.89

aneously

0.88

ricular

0.84

imester

0.80

 noon

0.80

 midnight

0.78

 graduation

0.76

enges

0.76

 lunch

0.73

rely

0.72

Activations Density 0.029%

patterns related to events happening immediately one after another

phrases indicating sequential events or occurrences

No Comments

No Known Activations