INDEX

Explanations

mentions of specific TV show episodes

oai_token-act-pair · gpt-3.5-turbo

references to specific television episodes or series

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

 trough

-0.73

 clasp

-0.68

TeX

-0.66

 notation

-0.66

 annex

-0.65

lig

-0.65

 monop

-0.63

 vest

-0.63

ulkan

-0.63

 motif

-0.63

POSITIVE LOGITS

 Showtime

1.10

HBO

1.08

 Dancing

1.07

 MSNBC

1.06

NBC

1.02

 Celebrity

1.01

CSI

0.99

 Survivor

0.99

 Bravo

0.98

 Comedy

0.96

Activations Density 0.166%

mentions of specific TV show episodes

references to specific television episodes or series

No Comments

No Known Activations