INDEX

Explanations

strings related to following or adhering to instructions, directions, or guidelines

oai_token-act-pair · gpt-3.5-turbo

references to following or adhering to concepts or rules

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

ãĥĩãĤ£

-0.75

IOR

-0.64

itary

-0.63

roxy

-0.63

pload

-0.62

mu

-0.62

azz

-0.61

ability

-0.61

being

-0.60

 usable

-0.59

POSITIVE LOGITS

 footsteps

1.53

 directions

1.26

 instructions

1.19

 path

1.04

 closely

1.03

 trail

1.01

 advice

0.99

 dictates

0.97

 trajectory

0.97

 footprints

0.96

Activations Density 0.137%

strings related to following or adhering to instructions, directions, or guidelines

references to following or adhering to concepts or rules

No Comments

No Known Activations