INDEX

Explanations

phrases related to visual art and object decoration

oai_token-act-pair · gpt-3.5-turbo

references to objects or concepts that are "painted."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

EMS

-0.87

fell

-0.74

cum

-0.72

FL

-0.71

PF

-0.69

Reason

-0.68

our

-0.67

hart

-0.67

conom

-0.67

ousing

-0.67

POSITIVE LOGITS

 mock

0.82

 wooden

0.81

 subur

0.78

 decoration

0.76

 painted

0.75

 pink

0.74

 ceramic

0.73

escription

0.71

 markings

0.71

 stainless

0.70

Activations Density 0.036%

phrases related to visual art and object decoration

references to objects or concepts that are "painted."

No Comments

No Known Activations