INDEX

Explanations

phrases or descriptions that are visually impressive or noteworthy

oai_token-act-pair · gpt-3.5-turbo

descriptors that indicate something is impressive or noteworthy

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

nav

-0.78

ruary

-0.78

OPE

-0.76

clerosis

-0.73

Ö¼

-0.71

iatrics

-0.70

nery

-0.69

cellence

-0.68

acca

-0.68

peria

-0.68

POSITIVE LOGITS

ly

1.07

LY

0.94

ively

0.89

tons

0.88

redients

0.77

NESS

0.75

 heights

0.71

 establishments

0.69

inventoryQuantity

0.68

 atmosp

0.68

Activations Density 0.099%

phrases or descriptions that are visually impressive or noteworthy

descriptors that indicate something is impressive or noteworthy

No Comments

No Known Activations