INDEX

Explanations

programming-related keywords and syntax

oai_token-act-pair · gpt-3.5-turbo

programming functions and database-related actions in code snippets

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

................................................................

-0.73

 gearing

-0.66

 spoiler

-0.62

 catast

-0.61

 spoilers

-0.60

................

-0.59

 troubles

-0.58

 shaping

-0.58

 choosing

-0.57

 lying

-0.57

POSITIVE LOGITS

();

1.54

("

1.35

((

1.34

('

1.28

({

1.24

(&

1.24

($

1.23

();

1.23

([

1.23

().

1.23

Activations Density 0.078%

programming-related keywords and syntax

programming functions and database-related actions in code snippets

No Comments

No Known Activations