INDEX

Explanations

occurrences of the word "in"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.8.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 furthe

-0.66

 servicing

-0.65

 lapt

-0.65

 foreseeable

-0.64

duino

-0.64

 everywhere

-0.63

-0.62

ooz

-0.61

ometimes

-0.61

 stretches

-0.61

POSITIVE LOGITS

 Proceedings

0.98

 Handbook

0.92

 Rats

0.80

:,

0.80

.:

0.78

 Supplement

0.73

 Gentleman

0.73

rint

0.71

 Proc

0.70

pp

0.69

Activations Density 0.253%

occurrences of the word "in"

No Comments

No Known Activations