INDEX

Explanations

terms related to news, events, and reporting

oai_token-act-pair · gpt-3.5-turbo

terms related to evidence and reporting in news 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-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

âĢķ

-0.78

Â¯

-0.77

––

-0.74

Ïī

-0.71

(<

-0.67

*.

-0.66

 EVERY

-0.65

("

-0.65

 latitude

-0.64

 every

-0.64

POSITIVE LOGITS

 Highlights

1.02

 Details

0.93

 spokesman

0.92

 spokeswoman

0.90

 Continued

0.88

 researcher

0.87

 spokesperson

0.86

 Updates

0.83

 analyst

0.83

 observers

0.81

Activations Density 0.280%

terms related to news, events, and reporting

terms related to evidence and reporting in news contexts

No Comments

No Known Activations