INDEX

Explanations

phrases indicating potential consequences or possibilities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "may" related to potential outcomes or hypotheses

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

ocracy

-0.87

ging

-0.82

 Fighter

-0.71

ament

-0.68

raged

-0.68

 Enforcement

-0.65

ãĥ¼ãĥ«

-0.64

brance

-0.64

zeb

-0.64

 Lifetime

-0.63

POSITIVE LOGITS

 derive

0.84

entimes

0.83

hap

0.83

 misunder

0.82

 inadvertently

0.82

 mistakenly

0.82

berra

0.82

 alternatively

0.81

onna

0.81

be

0.81

Activations Density 0.074%

phrases indicating potential consequences or possibilities

occurrences of the word "may" related to potential outcomes or hypotheses

No Comments

No Known Activations