INDEX

Explanations

pairs of related entities or concepts

oai_token-act-pair · gpt-3.5-turbo

references to pairs, especially in the context of people or 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

ãĤ´ãĥ³

-0.81

ãĤ¼ãĤ¦ãĤ¹

-0.67

taboola

-0.65

rf

-0.63

advertisement

-0.62

ãĤ¦

-0.60

UNE

-0.60

phi

-0.60

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.59

une

-0.58

POSITIVE LOGITS

 totaling

1.08

 apiece

0.97

 halves

0.82

 consecut

0.81

 sisters

0.80

 thirds

0.77

 identical

0.76

 simultaneously

0.76

 brothers

0.76

 finalists

0.75

Activations Density 0.574%

pairs of related entities or concepts

references to pairs, especially in the context of people or entities

No Comments

No Known Activations