INDEX

Explanations

the name "Eren" in various contexts

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of a specific name or term

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Bust

-0.72

 raises

-0.64

 pledges

-0.62

 Must

-0.61

 radius

-0.60

 grip

-0.60

 pledge

-0.59

XX

-0.58

 stock

-0.58

 bowl

-0.58

POSITIVE LOGITS

eren

4.65

ere

1.53

erent

1.50

erences

1.25

ERE

1.20

aren

1.17

eree

1.11

oren

1.04

eln

1.03

erer

1.02

Activations Density 0.009%

the name "Eren" in various contexts

repeated mentions of a specific name or term

No Comments

No Known Activations