INDEX

Explanations

sequences of repeated letters

oai_token-act-pair · gpt-3.5-turbo

expressions of excitement or surprise

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

ourse

-0.76

"]=>

-0.64

 contribut

-0.62

ãĥĹ

-0.61

 substitution

-0.60

IAL

-0.59

 vacated

-0.57

 Redux

-0.57

occup

-0.57

 Hasan

-0.57

POSITIVE LOGITS

mmmm

1.04

mmm

0.91

oooo

0.91

 kidding

0.87

ooo

0.87

ahah

0.84

hhhh

0.82

aaaa

0.79

hhh

0.79

!!!!!

0.78

Activations Density 0.355%

sequences of repeated letters

expressions of excitement or surprise

No Comments

No Known Activations