INDEX

Explanations

proper nouns, likely names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of people and entities

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

âĶĢâĶĢ

-0.84

etheless

-0.80

DCS

-0.71

BILITIES

-0.70

ournal

-0.70

imbabwe

-0.69

UTERS

-0.69

ailability

-0.68

ATIONAL

-0.68

ITIES

-0.68

POSITIVE LOGITS

hart

0.98

hoff

0.96

enberg

0.96

mann

0.91

meyer

0.91

gaard

0.90

stad

0.90

gren

0.90

burn

0.90

ley

0.88

Activations Density 0.223%

proper nouns, likely names of individuals

proper nouns, particularly names of people and entities

No Comments

No Known Activations