INDEX

Explanations

text related to technical data formatting or specifications

oai_token-act-pair · gpt-3.5-turbo

instances of "IN" followed by varying numerical indicators or references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

EStream

-0.74

uyomi

-0.70

 Brach

-0.70

ske

-0.67

liga

-0.65

catentry

-0.65

 grounds

-0.65

 Fired

-0.65

 BaseType

-0.65

 compr

-0.62

POSITIVE LOGITS

TERN

1.48

STRUCT

1.30

CLUD

1.21

CLUS

1.19

FORMATION

1.15

HER

1.15

CEPT

1.14

PUT

1.10

VERS

1.10

strument

1.08

Activations Density 0.009%

text related to technical data formatting or specifications

instances of "IN" followed by varying numerical indicators or references

No Comments

No Known Activations