INDEX

Explanations

words related to physical locations, especially those within a medical or institutional context

oai_token-act-pair · gpt-3.5-turbo

references to medical wards and their characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥĪ

-0.87

ãĥ¤

-0.87

ctory

-0.72

pmwiki

-0.72

igslist

-0.72

Hav

-0.72

kered

-0.71

ãĥ¬

-0.70

ISM

-0.69

itars

-0.69

POSITIVE LOGITS

robe

1.27

room

1.07

 ward

1.05

nton

0.89

masters

0.85

rooms

0.84

iencies

0.83

lings

0.81

 wards

0.79

abouts

0.77

Activations Density 0.013%

words related to physical locations, especially those within a medical or institutional context

references to medical wards and their characteristics

No Comments

No Known Activations