INDEX

Explanations

specific words preceded by 'the' or similar articles

oai_token-act-pair · gpt-3.5-turbo

terms related to scientific or philosophical concepts

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

 grab

-0.66

sed

-0.64

perty

-0.62

accompan

-0.62

colour

-0.59

issued

-0.59

 elig

-0.59

cit

-0.57

cki

-0.57

 blazing

-0.56

POSITIVE LOGITS

heim

0.79

neys

0.73

Henry

0.73

umbai

0.68

bidden

0.67

leans

0.66

 Awakens

0.65

Recipe

0.65

ignant

0.65

itudes

0.64

Activations Density 0.031%

specific words preceded by 'the' or similar articles

terms related to scientific or philosophical concepts

No Comments

No Known Activations