INDEX

Explanations

references to specific time points in the narrative

oai_token-act-pair · gpt-3.5-turbo

references to significant events or occurrences in a narrative structure

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dain

-0.77

="#

-0.71

 Tradable

-0.68

å§«

-0.65

pread

-0.64

ansk

-0.63

NYSE

-0.63

é¾įå¥ĳå£«

-0.62

 Answer

-0.61

hess

-0.60

POSITIVE LOGITS

 year

1.36

 ensuing

1.32

 same

1.28

 following

1.28

 subsequent

1.24

 month

1.21

 next

1.11

decade

1.09

 intervening

1.09

 succeeding

1.02

Activations Density 0.233%

references to specific time points in the narrative

references to significant events or occurrences in a narrative structure

No Comments

No Known Activations