INDEX

Explanations

references to specific events, people, or numerical data

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.10.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

 scrut

-0.77

 multic

-0.68

 redes

-0.68

semb

-0.68

%]

-0.67

PDF

-0.67

 accur

-0.64

authorized

-0.64

 expended

-0.63

athered

-0.63

POSITIVE LOGITS

>>

1.28

>>

1.00

 Honestly

0.80

Hey

0.80

Huh

0.80

 Yeah

0.76

 Fuck

0.76

 Anon

0.73

 Originally

0.73

Nig

0.73

Activations Density 0.154%

references to specific events, people, or numerical data

No Comments

No Known Activations