INDEX

Explanations

locations and time references

oai_token-act-pair · gpt-3.5-turbo

locations and references to specific places or settings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.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

 fundament

-0.61

auts

-0.55

 epit

-0.54

ollah

-0.54

Qi

-0.53

ãĥĥãĥĪ

-0.53

ãĥ«

-0.52

 bitters

-0.52

 mosaic

-0.52

 ensued

-0.52

POSITIVE LOGITS

 lookout

0.76

 awhile

0.75

ournals

0.70

adj

0.67

elight

0.66

 somewhere

0.65

 waiting

0.64

alert

0.64

 myself

0.64

 illegally

0.64

Activations Density 0.527%

locations and time references

locations and references to specific places or settings

No Comments

No Known Activations