INDEX

Explanations

words related to names or terms starting with "Lev"

oai_token-act-pair · gpt-3.5-turbo

references to the name "Lev"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ħĭ

-0.78

REDACTED

-0.69

PLA

-0.62

 Marcos

-0.60

 Kenobi

-0.59

ORGE

-0.59

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.59

reference

-0.59

Prin

-0.58

^^^^

-0.58

POSITIVE LOGITS

itic

1.45

Lev

1.25

itans

1.09

isions

1.09

iton

1.06

itate

1.04

itating

1.04

itan

1.02

inson

1.00

andowski

1.00

Activations Density 0.006%

words related to names or terms starting with "Lev"

references to the name "Lev"

No Comments

No Known Activations