INDEX

Explanations

references to geographical locations, particularly lakes

oai_token-act-pair · gpt-3.5-turbo

references to the word "Lake."

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

nir

-0.87

 tremend

-0.82

*/(

-0.81

AAAAAAAA

-0.77

imental

-0.74

endiary

-0.73

CLASSIFIED

-0.70

iotics

-0.70

nels

-0.70

binary

-0.69

POSITIVE LOGITS

 Erie

1.06

front

0.93

 trout

0.89

wood

0.89

 Superior

0.87

 Lake

0.85

 Nile

0.84

 Mead

0.84

 Lakes

0.82

bone

0.80

Activations Density 0.012%

references to geographical locations, particularly lakes

references to the word "Lake."

No Comments

No Known Activations