INDEX

Explanations

references to military ranks and titles, specifically focusing on the rank of Commander

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Commander" related to various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ramid

-0.67

ensions

-0.62

NRS

-0.61

employment

-0.61

 Marketable

-0.60

pher

-0.60

heres

-0.59

ether

-0.59

Gaza

-0.58

amina

-0.58

POSITIVE LOGITS

 Commander

1.36

utenant

0.90

 Shepard

0.84

 Specialist

0.82

chief

0.82

 commander

0.81

 Officer

0.80

 Command

0.79

EStream

0.78

 Duty

0.74

Activations Density 0.007%

references to military ranks and titles, specifically focusing on the rank of Commander

mentions of the term "Commander" related to various contexts

No Comments

No Known Activations