INDEX

Explanations

the word "Aust" or "Austrian" at various activation levels

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Aust" in various contexts, particularly related to Austro-Hungarian references

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

SPONSORED

-0.85

displayText

-0.76

FTWARE

-0.71

MLG

-0.69

EngineDebug

-0.67

WAYS

-0.67

TED

-0.66

PATH

-0.66

HEAD

-0.65

çīĪ

-0.64

POSITIVE LOGITS

erity

1.21

inite

1.08

rika

1.05

 Aust

1.04

rians

1.03

rator

1.01

ereo

1.01

rals

0.98

alis

0.97

rones

0.95

Activations Density 0.015%

the word "Aust" or "Austrian" at various activation levels

mentions of the word "Aust" in various contexts, particularly related to Austro-Hungarian references

No Comments

No Known Activations