INDEX

Explanations

terms related to conflict and disagreement

oai_token-act-pair · gpt-3.5-turbo

words associated with rancor and hostility

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

 Archdemon

-0.71

 Doctrine

-0.69

 Atlantis

-0.67

 Sorceress

-0.64

uyomi

-0.61

 Isles

-0.61

 Gravity

-0.61

 Fore

-0.61

etheless

-0.60

 Scotia

-0.59

POSITIVE LOGITS

 ranc

1.10

ÃŃa

0.86

yssey

0.85

iosity

0.85

 corrid

0.84

 ranch

0.81

aste

0.81

esse

0.80

iously

0.79

anth

0.78

Activations Density 0.007%

terms related to conflict and disagreement

words associated with rancor and hostility

No Comments

No Known Activations