INDEX

Explanations

phrases or words related to time sequences, particularly referring to events happening later

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "later."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

-0.66

ÂłÂłÂłÂłÂłÂłÂłÂł

-0.64

HY

-0.63

Pwr

-0.63

esome

-0.60

Merit

-0.60

cli

-0.59

ampa

-0.58

PORT

-0.56

CLOSE

-0.56

POSITIVE LOGITS

ally

1.24

ality

0.91

etheless

0.86

ificant

0.78

 than

0.77

iated

0.77

 generations

0.77

ons

0.76

aneously

0.76

 regretted

0.75

Activations Density 0.047%

phrases or words related to time sequences, particularly referring to events happening later

mentions of the word "later."

No Comments

No Known Activations