INDEX

Explanations

years in the late 2000s, particularly 2008

oai_token-act-pair · gpt-3.5-turbo

references to the year 2008 and its significant events

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

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

 ancest

-0.84

detail

-0.72

lihood

-0.68

ymes

-0.68

 Rooney

-0.63

ãĤ¤ãĥĪ

-0.60

ancy

-0.60

venge

-0.58

heid

-0.58

lier

-0.58

POSITIVE LOGITS

-'

0.73

irection

0.72

wark

0.71

å¹

0.71

iyah

0.68

 Reconstruction

0.67

finals

0.66

bit

0.65

yrinth

0.64

 onwards

0.63

Activations Density 0.027%

years in the late 2000s, particularly 2008

references to the year 2008 and its significant events

No Comments

No Known Activations