INDEX

Explanations

mentions of book titles and authors

oai_token-act-pair · gpt-3.5-turbo

titles of books and references to authors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 preferably

-0.66

 salsa

-0.66

 similar

-0.64

 staff

-0.63

 quir

-0.61

TN

-0.61

 ambulance

-0.61

 projectile

-0.60

 talented

-0.59

 closest

-0.59

POSITIVE LOGITS

 Revis

1.25

 Versus

1.19

 Strategies

1.16

 Handbook

1.15

 Patterns

1.13

 Matters

1.12

 Manifest

1.10

 Lessons

1.10

 Causes

1.09

 Expect

1.08

Activations Density 0.278%

mentions of book titles and authors

titles of books and references to authors

No Comments

No Known Activations