INDEX

Explanations

phrases related to decision-making and granting permissions

oai_token-act-pair · gpt-3.5-turbo

phrases related to rights and responsibilities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Ö¼

-0.76

zar

-0.75

sever

-0.74

minus

-0.70

albeit

-0.69

Ĥª

-0.67

nsic

-0.66

\\\\\\\\

-0.66

 immensely

-0.64

ĺ

-0.62

POSITIVE LOGITS

 respectively

0.92

 versus

0.92

or

0.80

 appropriate

0.79

vs

0.74

etc

0.74

 best

0.73

 each

0.72

 prioritize

0.72

 priorit

0.71

Activations Density 0.418%

phrases related to decision-making and granting permissions

phrases related to rights and responsibilities

No Comments

No Known Activations