INDEX

Explanations

adjectives describing actions or situations that are negatively perceived

oai_token-act-pair · gpt-3.5-turbo

terms related to stopping or blaming in various contexts

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

ultane

-0.82

ãĥİ

-0.72

 teasp

-0.66

Interstitial

-0.65

 exting

-0.64

=~=~

-0.62

ignt

-0.62

irez

-0.61

FORMATION

-0.60

otation

-0.59

POSITIVE LOGITS

 Canaver

0.91

 libertarians

0.79

 Krugman

0.78

 Bezos

0.77

 Turnbull

0.76

 Redmond

0.76

 economists

0.74

 Elon

0.74

 liberals

0.73

 Giuliani

0.73

Activations Density 0.598%

adjectives describing actions or situations that are negatively perceived

terms related to stopping or blaming in various contexts

No Comments

No Known Activations