INDEX

Explanations

instances of the word "real" emphasizing authenticity or significance

oai_token-act-pair · gpt-3.5-turbo

instances of the word "real" and its variations that imply authenticity or significance

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

ulations

-0.86

anwhile

-0.85

poons

-0.84

artments

-0.82

runs

-0.78

xual

-0.77

moil

-0.77

utics

-0.77

encies

-0.76

lations

-0.76

POSITIVE LOGITS

ignment

1.02

 contender

0.99

 thing

0.90

 indication

0.88

 boon

0.85

 headache

0.84

 killer

0.83

 believer

0.82

 indicator

0.81

 testament

0.81

Activations Density 0.071%

instances of the word "real" emphasizing authenticity or significance

instances of the word "real" and its variations that imply authenticity or significance

No Comments

No Known Activations