INDEX

Explanations

words related to a specific brand or company, containing the letters "ire"

oai_token-act-pair · gpt-3.5-turbo

words or phrases pertaining to specific product features or brand elements

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

ļé

-0.86

ĸļ

-0.81

romeda

-0.81

 carbohyd

-0.79

¬¼

-0.77

sylv

-0.73

acca

-0.71

 cellul

-0.70

achusetts

-0.69

ockets

-0.69

POSITIVE LOGITS

nces

1.04

lli

1.00

lessly

0.91

ly

0.88

lette

0.83

lla

0.82

ndra

0.81

ussia

0.80

zza

0.79

cia

0.77

Activations Density 0.027%

words related to a specific brand or company, containing the letters "ire"

words or phrases pertaining to specific product features or brand elements

No Comments

No Known Activations