INDEX

Explanations

mentions of actions or events from the past

oai_token-act-pair · gpt-3.5-turbo

references to events or actions that have occurred in the past

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

hat

-0.72

 Franch

-0.69

wagon

-0.68

starting

-0.68

oya

-0.68

MSN

-0.67

hyde

-0.66

atism

-0.65

NEY

-0.65

ODY

-0.64

POSITIVE LOGITS

ebin

1.42

 tense

1.28

decade

1.27

few

1.13

 fortnight

1.10

 couple

1.06

 month

1.05

 week

1.00

orate

0.99

 century

0.96

Activations Density 0.030%

mentions of actions or events from the past

references to events or actions that have occurred in the past

No Comments

No Known Activations