INDEX

Explanations

the word "series"

oai_token-act-pair · gpt-3.5-turbo

the phrase "a series of" followed by various nouns or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ability

-0.71

steen

-0.67

stadt

-0.66

dad

-0.65

gow

-0.64

Introduced

-0.63

tera

-0.62

vale

-0.62

heed

-0.62

liest

-0.61

POSITIVE LOGITS

 events

0.76

 releases

0.74

 alternating

0.73

 sorts

0.73

 course

0.64

 successive

0.63

 consecutive

0.62

course

0.62

 articles

0.61

 posts

0.61

Activations Density 0.074%

the word "series"

the phrase "a series of" followed by various nouns or concepts

No Comments

No Known Activations