INDEX

Explanations

the word "probably"

oai_token-act-pair · gpt-3.5-turbo

the word "probably" indicating uncertainty or speculation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

rients

-0.60

 Held

-0.59

];

-0.58

ento

-0.57

 Seeking

-0.57

ins

-0.56

powers

-0.56

iegel

-0.56

ignant

-0.55

ween

-0.55

POSITIVE LOGITS

 probably

3.20

probably

2.63

 doubtless

2.36

 undoubtedly

2.13

 definitely

2.08

 Probably

2.06

 likely

2.03

 certainly

1.98

Probably

1.91

 presumably

1.91

Activations Density 0.022%

the word "probably"

the word "probably" indicating uncertainty or speculation

No Comments

No Known Activations