INDEX

Explanations

references to the term "Brothers"

oai_token-act-pair · gpt-3.5-turbo

references to specific "Brothers" entities across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

dec

-0.73

 retention

-0.72

 cavity

-0.72

CET

-0.68

 dent

-0.68

 biod

-0.67

 policy

-0.66

 privacy

-0.65

 gloss

-0.65

rid

-0.65

POSITIVE LOGITS

 Brothers

4.19

 Bros

2.27

 brothers

2.21

Brother

2.17

 Brother

2.02

 Sisters

1.94

brother

1.84

 sisters

1.61

 Brotherhood

1.43

 Sister

1.30

Activations Density 0.007%

references to the term "Brothers"

references to specific "Brothers" entities across various contexts

No Comments

No Known Activations