INDEX

Explanations

adjectives and nouns related to political stances or policies

oai_token-act-pair · gpt-3.5-turbo

evaluative phrases that categorize policies as beneficial or detrimental

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

videos

-0.81

RTX

-0.78

views

-0.78

ULTS

-0.77

lees

-0.75

stories

-0.74

images

-0.71

notes

-0.71

 backgrounds

-0.70

itely

-0.69

POSITIVE LOGITS

 boon

1.24

 deterrent

1.16

 subsidy

1.15

 safeguard

1.09

 bailout

1.08

 cornerstone

1.07

 giveaway

1.06

 tool

1.06

 carrot

1.04

 Trojan

1.04

Activations Density 0.225%

adjectives and nouns related to political stances or policies

evaluative phrases that categorize policies as beneficial or detrimental

No Comments

No Known Activations