INDEX

Explanations

words related to things that are hidden or unnoticed

oai_token-act-pair · gpt-3.5-turbo

terms related to unreleased content and themes of unavailability or incompleteness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

uyomi

-0.77

 Labrador

-0.68

Gab

-0.68

 Higgins

-0.68

enegger

-0.67

hetti

-0.67

anwhile

-0.66

ospace

-0.63

=-=-=-=-=-=-=-=-

-0.63

ea

-0.63

POSITIVE LOGITS

leased

1.12

achable

1.10

 unre

0.93

itled

0.92

ached

0.90

ason

0.87

ritten

0.85

agan

0.85

ported

0.84

bled

0.82

Activations Density 0.008%

words related to things that are hidden or unnoticed

terms related to unreleased content and themes of unavailability or incompleteness

No Comments

No Known Activations