INDEX

Explanations

time-related phrases, such as "a decade ago" or "last year."

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 twitch

-0.64

aw

-0.58

abies

-0.58

press

-0.57

ruck

-0.57

abase

-0.57

ila

-0.57

enance

-0.56

 Chandler

-0.55

amazon

-0.55

POSITIVE LOGITS

 lows

0.78

 predec

0.77

 highs

0.72

 estim

0.71

iator

0.70

 glim

0.70

 eras

0.69

 selves

0.68

 predecessors

0.68

verty

0.64

Activations Density 0.192%

time-related phrases, such as "a decade ago" or "last year."

No Comments

No Known Activations