INDEX

Explanations

actions or situations where something ends up in a particular state or place

oai_token-act-pair · gpt-3.5-turbo

variations of the word "wound."

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

alty

-0.86

poons

-0.80

acus

-0.71

 chefs

-0.71

venants

-0.68

bour

-0.65

 principals

-0.65

 juveniles

-0.65

 chef

-0.64

cale

-0.64

POSITIVE LOGITS

stead

0.80

bare

0.75

itial

0.75

 winding

0.75

strap

0.74

ework

0.72

ners

0.72

terson

0.72

neck

0.71

osc

0.70

Activations Density 0.027%

actions or situations where something ends up in a particular state or place

variations of the word "wound."

No Comments

No Known Activations