INDEX

Explanations

historical references or events

oai_token-act-pair · gpt-3.5-turbo

the repeated mention of the word "back" in historical 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

haz

-0.64

 understatement

-0.63

tyr

-0.63

ellen

-0.62

entric

-0.62

ittee

-0.62

 Sorceress

-0.60

izo

-0.58

ittees

-0.57

 Corpse

-0.56

POSITIVE LOGITS

dated

1.17

fired

1.12

fires

1.07

packing

1.06

stab

1.04

packs

0.97

stage

0.97

dating

0.96

gam

0.93

country

0.93

Activations Density 0.027%

historical references or events

the repeated mention of the word "back" in historical contexts

No Comments

No Known Activations