INDEX

Explanations

references to large or oversized objects

oai_token-act-pair · gpt-3.5-turbo

references to "giant" entities or characters

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

yrinth

-0.91

anwhile

-0.83

qi

-0.77

endment

-0.77

earchers

-0.76

imester

-0.73

eligible

-0.73

rences

-0.72

informed

-0.72

iggins

-0.72

POSITIVE LOGITS

 squid

1.03

 chunk

0.85

 sized

0.83

 gorilla

0.81

 leap

0.79

 monster

0.79

gest

0.78

 bould

0.77

 chunks

0.76

 slug

0.76

Activations Density 0.020%

references to large or oversized objects

references to "giant" entities or characters

No Comments

No Known Activations