INDEX

Explanations

ranges of values or quantities

oai_token-act-pair · gpt-3.5-turbo

ranges or quantities expressed numerically

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

aign

-0.64

hemat

-0.63

ById

-0.58

BALL

-0.57

inventory

-0.57

vir

-0.56

dain

-0.55

emetery

-0.55

birds

-0.54

calling

-0.54

POSITIVE LOGITS

ggles

1.06

pless

0.89

ppers

0.80

pload

0.79

kes

0.76

kus

0.74

lling

0.70

 compensate

0.69

asted

0.69

mes

0.68

Activations Density 0.065%

ranges of values or quantities

ranges or quantities expressed numerically

No Comments

No Known Activations