INDEX

Explanations

proper nouns, especially focusing on last names

oai_token-act-pair · gpt-3.5-turbo

names and terms related to specific individuals, particularly those associated with notable figures or events

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

lined

-0.73

accompan

-0.71

 thinner

-0.68

 differentiated

-0.65

oding

-0.63

rera

-0.63

ŃĶ

-0.63

uscript

-0.62

erick

-0.60

 cryst

-0.60

POSITIVE LOGITS

enger

0.96

mann

0.92

iev

0.80

minster

0.79

nut

0.77

iance

0.77

hent

0.77

ynski

0.76

utz

0.76

burgh

0.76

Activations Density 0.020%

proper nouns, especially focusing on last names

names and terms related to specific individuals, particularly those associated with notable figures or events

No Comments

No Known Activations