INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to geographical locations and historical references

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

CLOSE

-0.73

udeb

-0.66

 tenants

-0.64

ESH

-0.62

falls

-0.61

holding

-0.60

 disg

-0.59

 rolling

-0.59

ppings

-0.58

CVE

-0.58

POSITIVE LOGITS

kefeller

0.97

uthor

0.91

anyahu

0.87

hower

0.85

zsche

0.82

ource

0.79

heid

0.75

jriwal

0.73

utenant

0.73

tainment

0.72

Activations Density 0.164%

proper nouns

proper nouns related to geographical locations and historical references

No Comments

No Known Activations