INDEX

Explanations

mentions of meetings at specific locations

oai_token-act-pair · gpt-3.5-turbo

names of political events and related locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 untold

-0.82

 dearly

-0.68

ãĥĥãĥī

-0.67

etheless

-0.65

govtrack

-0.63

abled

-0.63

theless

-0.62

 altogether

-0.60

 gladly

-0.60

otally

-0.58

POSITIVE LOGITS

 REUTERS

1.15

 Thomson

0.89

 Reuters

0.87

 during

0.83

 October

0.78

 November

0.78

 February

0.77

 Â©

0.77

EPA

0.77

 April

0.76

Activations Density 0.219%

mentions of meetings at specific locations

names of political events and related locations

No Comments

No Known Activations