INDEX

Explanations

scientific research-related terms and findings

oai_token-act-pair · gpt-3.5-turbo

statements about surprising scientific results and their implications

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

 hating

-0.77

 Exile

-0.76

Rated

-0.74

 Theft

-0.70

 Seasons

-0.70

 Sovereign

-0.68

 Destination

-0.67

 Expend

-0.66

 Municipal

-0.65

 despise

-0.65

POSITIVE LOGITS

 intriguing

1.23

 hypothesis

1.13

 breakthrough

1.12

 implications

1.09

 exciting

1.09

 hypothes

1.08

 tantal

1.08

 hypotheses

1.07

 promising

1.07

 preliminary

1.06

Activations Density 0.424%

scientific research-related terms and findings

statements about surprising scientific results and their implications

No Comments

No Known Activations