INDEX

Explanations

terms related to media coverage or reporting

oai_token-act-pair · gpt-3.5-turbo

instances of the word "coverage," particularly in contexts related to media and reporting

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

 Sabha

-0.73

 slightest

-0.72

gery

-0.70

geries

-0.69

rir

-0.67

entle

-0.67

pert

-0.63

mine

-0.62

warm

-0.62

 restless

-0.61

POSITIVE LOGITS

 Coverage

1.11

 coverage

1.05

cov

0.85

verage

0.84

krit

0.78

oided

0.77

irtual

0.77

orship

0.73

anyahu

0.73

 COVER

0.72

Activations Density 0.023%

terms related to media coverage or reporting

instances of the word "coverage," particularly in contexts related to media and reporting

No Comments

No Known Activations