INDEX

Explanations

foreign locations or cultural references

oai_token-act-pair · gpt-3.5-turbo

vowel-heavy words or sequences in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

kees

-0.77

ĸļ

-0.70

ONSORED

-0.69

gnu

-0.67

ukemia

-0.66

 leukemia

-0.65

Ïī

-0.65

sbm

-0.65

 unborn

-0.64

uckland

-0.64

POSITIVE LOGITS

 Lauder

0.99

 Sant

0.91

olini

0.78

ateral

0.77

Ã±

0.76

angelo

0.76

 Rican

0.75

inas

0.74

Ã±o

0.73

é¾įå

0.72

Activations Density 0.209%

foreign locations or cultural references

vowel-heavy words or sequences in text

No Comments

No Known Activations