INDEX

Explanations

terms related to a specific location, "Gat."

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or geographic names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Nile

-0.83

opian

-0.70

vul

-0.69

ering

-0.69

vol

-0.66

ebook

-0.65

 dear

-0.63

agra

-0.63

LIB

-0.62

 sidel

-0.60

POSITIVE LOGITS

ãĥģ

0.94

ettle

0.77

herer

0.73

atory

0.73

Â¯Â¯Â¯Â¯

0.73

ãĥĸ

0.73

ODY

0.73

enna

0.72

bands

0.71

 âĸº

0.71

Activations Density 0.050%

terms related to a specific location, "Gat."

references to specific locations or geographic names

No Comments

No Known Activations