INDEX

Explanations

references to elections and political candidacies

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

illions

-0.79

 Collider

-0.74

 OPEC

-0.73

 unanim

-0.72

facts

-0.69

 experts

-0.67

 theorists

-0.66

olars

-0.65

 warheads

-0.65

 iceberg

-0.65

POSITIVE LOGITS

 district

0.96

 incumbent

0.89

 lieutenant

0.88

 unsuccessfully

0.88

District

0.87

 seat

0.85

 challenger

0.83

 District

0.82

 township

0.79

iffe

0.79

Activations Density 1.708%

references to elections and political candidacies

No Comments

No Known Activations