INDEX

Explanations

phrases indicating stories or narratives

oai_token-act-pair · gpt-3.5-turbo

instances of the word "tale" and variations thereof

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

erate

-0.96

ividual

-0.80

bledon

-0.78

atever

-0.74

ournal

-0.73

ulty

-0.73

apsed

-0.73

eeks

-0.73

ussy

-0.73

uters

-0.72

POSITIVE LOGITS

 tales

1.27

 tale

1.23

tale

1.05

 Tale

1.01

 Tales

1.00

 Reincarn

0.91

tell

0.86

telling

0.77

Narr

0.75

crow

0.74

Activations Density 0.027%

phrases indicating stories or narratives

instances of the word "tale" and variations thereof

No Comments

No Known Activations