INDEX

Explanations

words related to disguise

oai_token-act-pair · gpt-3.5-turbo

instances of the word "gu" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Spectrum

-0.81

ŃĶ

-0.75

hower

-0.73

 cycle

-0.68

 Parables

-0.65

 spect

-0.64

 exting

-0.64

 eleph

-0.63

 Patriot

-0.61

 Triumph

-0.61

POSITIVE LOGITS

arant

1.12

idelines

1.07

arding

1.04

inea

0.95

arded

0.93

vernment

0.92

errilla

0.92

cci

0.91

ossip

0.90

atem

0.90

Activations Density 0.008%

words related to disguise

instances of the word "gu" in various contexts

No Comments

No Known Activations