INDEX

Explanations

words related to achievements or awards

oai_token-act-pair · gpt-3.5-turbo

references to awards or achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

stairs

-0.72

 Dynamo

-0.72

tons

-0.70

 seiz

-0.69

quickShipAvailable

-0.68

 =================================================================

-0.68

ENSE

-0.66

 orcs

-0.66

agin

-0.63

 correspondence

-0.63

POSITIVE LOGITS

xit

0.69

Apply

0.67

ãĤ¹ãĥĪ

0.66

 Converted

0.65

 Protection

0.61

ģ«

0.60

 Rasmussen

0.60

 Survey

0.60

 Banks

0.60

 Summary

0.58

Activations Density 0.000%

words related to achievements or awards

references to awards or achievements

No Comments

No Known Activations