INDEX

Explanations

phrases related to events leading up to specific occurrences

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a temporal sequence of events leading up to a significant occurrence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

icular

-0.69

owsky

-0.68

thr

-0.66

 scapego

-0.65

ums

-0.64

GA

-0.61

inav

-0.60

mens

-0.59

ku

-0.59

inen

-0.58

POSITIVE LOGITS

actionDate

1.01

stage

0.88

imester

0.86

uberty

0.77

ounter

0.77

dating

0.77

ãĥ©ãĥ³

0.72

isode

0.72

 thereto

0.71

stairs

0.69

Activations Density 0.050%

phrases related to events leading up to specific occurrences

phrases indicating a temporal sequence of events leading up to a significant occurrence

No Comments

No Known Activations