INDEX

Explanations

mentions of authors or writers in a positive context

oai_token-act-pair · gpt-3.5-turbo

references to authors and the act of writing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 DHCP

-0.82

 NOAA

-0.82

 FEMA

-0.81

 Berm

-0.79

DHS

-0.74

onomous

-0.73

DOE

-0.71

assis

-0.71

 Bash

-0.70

 VMware

-0.69

POSITIVE LOGITS

wright

0.97

writer

0.89

writers

0.88

acters

0.86

 muse

0.86

writing

0.85

ruary

0.85

igraph

0.82

 originals

0.80

 classics

0.79

Activations Density 0.193%

mentions of authors or writers in a positive context

references to authors and the act of writing

No Comments

No Known Activations