INDEX

Explanations

phrases indicating support or permission in various contexts

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

Lear

-0.72

ertodd

-0.71

 Survive

-0.70

 wandered

-0.69

AppData

-0.68

 wander

-0.67

 Grow

-0.67

isan

-0.66

 Conquer

-0.66

orem

-0.66

POSITIVE LOGITS

 awarding

1.07

 awarded

1.03

 sanction

0.97

 honored

0.95

 dispatched

0.93

 appoint

0.93

 sanctioned

0.93

 telev

0.93

 approve

0.91

 rescind

0.91

Activations Density 2.037%

phrases indicating support or permission in various contexts

No Comments

No Known Activations