INDEX

Explanations

references to anatomical "back" parts or directions, either anatomical structures or spatial positions

oai_token-act-pair · gpt-3.5-turbo

references to the term "back" in various contexts

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

iji

-0.71

ifix

-0.69

itizen

-0.68

entric

-0.67

Ń·

-0.66

izo

-0.66

 Parenthood

-0.66

aukee

-0.65

Cola

-0.64

ourning

-0.64

POSITIVE LOGITS

side

1.26

woods

1.11

packs

1.08

GROUND

0.99

seat

0.99

stab

0.97

hand

0.96

stories

0.96

drops

0.95

door

0.94

Activations Density 0.032%

references to anatomical "back" parts or directions, either anatomical structures or spatial positions

references to the term "back" in various contexts

No Comments

No Known Activations