INDEX

Explanations

the word "To" in various contexts and its frequency of use

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

 intrusive

-0.64

 dece

-0.64

 intentions

-0.63

 exceptions

-0.63

 proposals

-0.61

 regards

-0.60

 flagged

-0.60

 forthcoming

-0.60

 vulnerabilities

-0.58

 signatures

-0.58

POSITIVE LOGITS

ilet

1.71

pping

1.30

ilers

1.21

pped

1.15

asted

1.14

ffee

1.10

ppings

1.08

pper

1.05

ppers

1.05

asts

1.03

Activations Density 0.028%

the word "To" in various contexts and its frequency of use

No Comments

No Known Activations