INDEX

Explanations

the acronym "IM" or words related to management

oai_token-act-pair · gpt-3.5-turbo

references to the institution "IM" and related entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

yards

-0.82

 Brach

-0.69

 Thames

-0.69

halla

-0.69

 Borough

-0.66

 Vale

-0.64

 Chrys

-0.62

 greens

-0.62

velt

-0.62

Downloadha

-0.62

POSITIVE LOGITS

PLE

1.13

PLIC

1.07

MY

1.05

MED

0.99

pson

0.99

ITED

0.96

aterial

0.96

orph

0.96

ple

0.93

HO

0.93

Activations Density 0.006%

the acronym "IM" or words related to management

references to the institution "IM" and related entities

No Comments

No Known Activations