INDEX

Explanations

references to specific characters or people by their names

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to a specific identity or person

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 bilateral

-0.61

 jaws

-0.60

 Frey

-0.60

 banging

-0.58

 tram

-0.57

 mineral

-0.57

 heads

-0.56

 growing

-0.56

RAD

-0.56

 Vald

-0.56

POSITIVE LOGITS

lest

1.16

eller

1.11

leground

1.10

escription

1.07

lers

1.02

elf

1.01

irection

0.97

olesc

0.95

icter

0.95

idd

0.94

Activations Density 0.017%

references to specific characters or people by their names

references to a specific identity or person

No Comments

No Known Activations