INDEX

Explanations

mentions of specific locations or names

oai_token-act-pair · gpt-3.5-turbo

words containing the letters "nt" and "iant" 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 @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

aciously

-0.67

hammer

-0.66

cano

-0.65

perty

-0.63

bench

-0.59

 bait

-0.58

 downward

-0.57

 challeng

-0.55

mun

-0.55

STON

-0.55

POSITIVE LOGITS

1.11

ech

1.10

ean

1.10

yre

1.04

rek

1.02

oday

1.02

rix

1.00

own

0.97

ez

0.96

heastern

0.93

Activations Density 0.083%

mentions of specific locations or names

words containing the letters "nt" and "iant" in various contexts

No Comments

No Known Activations