INDEX

Explanations

information about quantities or limits

oai_token-act-pair · gpt-3.5-turbo

phrases indicating maximum limits or measurements

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

brook

-0.65

hare

-0.65

birds

-0.65

went

-0.63

NEY

-0.61

Moving

-0.61

rats

-0.60

èĢħ

-0.60

Chair

-0.59

Terror

-0.59

POSITIVE LOGITS

pload

0.95

ilings

0.77

 compensate

0.72

 date

0.71

asters

0.70

capacity

0.70

undreds

0.69

abyte

0.68

perty

0.68

undred

0.68

Activations Density 0.040%

information about quantities or limits

phrases indicating maximum limits or measurements

No Comments

No Known Activations