INDEX

Explanations

references to specific names, likely focusing on surnames

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to characters or entities in a specific context

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 cele

-0.83

 dent

-0.78

nai

-0.74

 hospitalized

-0.72

amy

-0.69

 herpes

-0.69

 paramedics

-0.68

 phot

-0.68

pus

-0.67

 cath

-0.66

POSITIVE LOGITS

 Frog

3.10

 Horn

2.05

 Blossom

1.45

 Crow

1.45

 Cotton

1.31

 Crusader

1.18

Fog

1.18

 Worm

1.17

 Rabbit

1.17

Egg

1.14

Activations Density 0.024%

references to specific names, likely focusing on surnames

proper nouns related to characters or entities in a specific context

No Comments

No Known Activations