INDEX

Explanations

phrases related to a specific film or book title

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and sentence boundaries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

uca

-0.74

 undocumented

-0.63

 unexplained

-0.61

 Viktor

-0.60

 scrap

-0.59

 Torch

-0.59

 unstoppable

-0.58

 Neville

-0.58

 elect

-0.58

 trusted

-0.56

POSITIVE LOGITS

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

1.06

ÂłÂłÂłÂł

1.04

SPONSORED

1.03

Each

1.01

Meanwhile

1.00

ccording

0.98

Advertisement

0.97

Similarly

0.96

ÂłÂłÂłÂłÂłÂłÂłÂł

0.95

Thus

0.95

Activations Density 0.910%

phrases related to a specific film or book title

punctuation marks and sentence boundaries

No Comments

No Known Activations