INDEX

Explanations

statements from computer programming code

oai_token-act-pair · gpt-3.5-turbo

programming-related logic and flow control structures

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

£ı

-0.62

 shed

-0.59

 disg

-0.58

opian

-0.58

anto

-0.57

ansk

-0.57

 commun

-0.56

ingen

-0.56

ittle

-0.55

ãĥ³ãĤ¸

-0.54

POSITIVE LOGITS

!=

1.52

==

1.50

==

1.41

>=

1.33

===

1.32

&&

1.30

<=

1.29

eq

1.11

||

1.05

){

1.02

Activations Density 0.083%

statements from computer programming code

programming-related logic and flow control structures

No Comments

No Known Activations