INDEX

Explanations

phrases related to communication and reporting

oai_token-act-pair · gpt-3.5-turbo

references to requests and communications regarding corrections or changes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 brutally

-0.76

 Whilst

-0.76

 ageing

-0.75

 whilst

-0.75

 brilliantly

-0.74

 fashionable

-0.74

isable

-0.71

 glor

-0.71

 murdering

-0.69

 ruth

-0.69

POSITIVE LOGITS

 spokeswoman

1.12

 commissioners

1.09

 Grassley

1.04

HHS

1.03

 attorneys

1.02

FEC

0.99

 spokesman

0.98

 McCabe

0.96

 supervisors

0.94

DOJ

0.94

Activations Density 0.854%

phrases related to communication and reporting

references to requests and communications regarding corrections or changes

No Comments

No Known Activations