INDEX

Explanations

words related to buildings, especially religious or historical ones such as monasteries, castles, and palaces

oai_token-act-pair · gpt-3.5-turbo

locations associated with religious or historical structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãģĹ

-0.68

etitive

-0.66

 Causes

-0.65

mitt

-0.65

 Voting

-0.62

ACTED

-0.61

acted

-0.60

Americans

-0.60

 Injury

-0.59

 Results

-0.59

POSITIVE LOGITS

 annex

1.02

 chapel

0.99

 decoration

0.92

 Dwell

0.90

 courtyard

0.89

 tower

0.88

 keeper

0.86

 dwelling

0.85

 palace

0.84

 complexes

0.82

Activations Density 0.057%

words related to buildings, especially religious or historical ones such as monasteries, castles, and palaces

locations associated with religious or historical structures

No Comments

No Known Activations