INDEX

Explanations

phrases indicating time periods, predominantly focusing on the current year

oai_token-act-pair · gpt-3.5-turbo

the word "this" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

oller

-0.69

gans

-0.69

rikes

-0.64

adle

-0.63

ansk

-0.62

quart

-0.62

uble

-0.61

facts

-0.61

MSN

-0.61

atos

-0.60

POSITIVE LOGITS

 year

1.37

 week

1.36

 semester

1.20

 weekend

1.19

 season

1.17

 month

1.17

 morning

1.13

 summer

1.11

 century

1.09

 afternoon

1.09

Activations Density 0.075%

phrases indicating time periods, predominantly focusing on the current year

the word "this" in various contexts

No Comments

No Known Activations