INDEX

Explanations

terms related to collisions or impacts

oai_token-act-pair · gpt-3.5-turbo

references to various types of collisions

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

estate

-0.84

ifter

-0.81

endment

-0.80

ibaba

-0.80

zel

-0.79

iatric

-0.77

vation

-0.77

azy

-0.77

secut

-0.76

iciency

-0.76

POSITIVE LOGITS

 collisions

0.80

SHIP

0.78

lag

0.75

engers

0.75

halla

0.73

 Franch

0.72

 collision

0.72

ships

0.71

 shenan

0.69

 Brend

0.69

Activations Density 0.017%

terms related to collisions or impacts

references to various types of collisions

No Comments

No Known Activations