INDEX

Explanations

mentions of sending troops or forces to various locations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to" and other phrases indicating sending or dispatching

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ById

-0.76

tions

-0.66

identally

-0.65

erg

-0.64

Mods

-0.62

ech

-0.61

walking

-0.60

ŃĶ

-0.60

 Signs

-0.60

cart

-0.60

POSITIVE LOGITS

 investigate

0.88

 oversee

0.83

 infiltrate

0.81

ongyang

0.79

 reinforce

0.77

 Pyongyang

0.77

 orbit

0.75

 Catalonia

0.75

 Fukushima

0.73

 Honduras

0.72

Activations Density 0.176%

mentions of sending troops or forces to various locations

instances of the word "to" and other phrases indicating sending or dispatching

No Comments

No Known Activations